We are now screening, staging and intervening first degree relatives of patients who have Type I diabetes for anti-islet cell antibodies to detect those subjects who have a moderate risk of developing diabetes in the future; further determination of the risk for Type I diabetes by examining the first phase of insulin secretion during an IV glucose tolerance test. An algorithm based on historical data will then be used to select those who have a greater than 50% risk of developing Type I diabetes; risk subjects will be randomized into an intervention study. Those with a 5 year risk between 25 and 50% are currently being randomized into a blinded trial of po insulin.